Research map / Label-efficient and robust learning / Recognition and long-tail
How Can I See My Future? FvTraj: Using First-person View for Pedestrian Trajectory Prediction
Huikun Bi, Ruisi Zhang, Tianlu Mao, Zhigang Deng, Zhaoqi Wang
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| add first-person view encoder and view-aware attention (ablation) | FvTraj · ETH/UCY average | ADE: 0.45 -> 0.41 | Improved |
| add first-person view encoder and view-aware attention (ablation) | FvTraj · ETH/UCY average | FDE: 0.94 -> 0.85 | Improved |
| add social-aware attention module (ablation) | FvTraj · ETH/UCY average | ADE: 0.45 -> 0.41 | Improved |
| add social-aware attention module (ablation) | FvTraj · ETH/UCY average | FDE: 0.91 -> 0.85 | Improved |
| add first-person view information (ablation) | FvTraj · HOTEL | ADE: 0.42 -> 0.28 | Improved |
| add first-person view information (ablation) | FvTraj · ETH | ADE: 0.58 -> 0.56 | Improved |
| add first-person view information (ablation) | FvTraj · ZARA1 | ADE: 0.37 -> 0.37 | No clear effect |
| add first-person view information (ablation) | FvTraj · ZARA2 | ADE: 0.33 -> 0.32 | No clear effect |
| increase variety-loss samples K from 1 to 20 (ablation) | FvTraj · ETH/UCY average | ADE: 0.47 -> 0.41 | Improved |
| increase variety-loss samples K from 1 to 20 (ablation) | FvTraj · ETH/UCY average | FDE: 0.96 -> 0.85 | Improved |
| increase variety-loss training from 1-1 to 5-20 (ablation) | FvTraj · ETH/UCY average | ADE: 0.50 -> 0.41 | Improved |
| use first-person view images instead of top-down view images | FvTraj vs Sophie/Social-BiGAT · ETH/UCY average | ADE: 0.54 -> 0.41 | Improved |
Closest papers by meaning
- Peeking Into the Future: Predicting Future Person Activities and Locations in Videos (CVPR 2019 · similarity 0.85)
- Forecasting Human Trajectory from Scene History (NeurIPS 2022 · similarity 0.85)
- GigaTraj: Predicting Long-term Trajectories of Hundreds of Pedestrians in Gigapixel Complex Scenes (CVPR 2024 · similarity 0.85)
- SocialCVAE: Predicting Pedestrian Trajectory via Interaction Conditioned Latents (AAAI 2024 · similarity 0.84)
- Social-Transmotion: Promptable Human Trajectory Prediction (ICLR 2024 · similarity 0.84)
- TrajCLIP: Pedestrian trajectory prediction method using contrastive learning and idempotent networks (NeurIPS 2024 · similarity 0.83)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
